In oxford-pharmacoepi/CodelistGenerator: Identify Relevant Clinical Codes and Evaluate Their Use

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%",
  message=FALSE, 
  warning=FALSE 
)

CodelistGenerator

Installation

You can install CodelistGenerator from CRAN

install.packages("CodelistGenerator")

Or you can also install the development version of CodelistGenerator

install.packages("remotes")
remotes::install_github("darwin-eu/CodelistGenerator")

Example usage

library(dplyr)
library(CDMConnector)
library(CodelistGenerator)

For this example we'll use the Eunomia dataset (which only contains a subset of the OMOP CDM vocabularies)

db <- DBI::dbConnect(duckdb::duckdb(), dbdir = eunomia_dir())
cdm <- cdm_from_con(db, cdm_schema = "main", write_schema = c(prefix = "cg_", schema = "main"))

Exploring the OMOP CDM Vocabulary tables

OMOP CDM vocabularies are frequently updated, and we can identify the version of the vocabulary of our Eunomia data

getVocabVersion(cdm = cdm)

CodelistGenerator provides various other functions to explore the vocabulary tables. For example, we can see the the different concept classes of standard concepts used for drugs

getConceptClassId(cdm,
                  standardConcept = "Standard",
                  domain = "Drug")

Vocabulary based codelists using CodelistGenerator

CodelistGenerator provides functions to extract code lists based on vocabulary hierarchies. One example is `getDrugIngredientCodes, which we can use, for example, to get all the concept IDs used to represent aspirin.

getDrugIngredientCodes(cdm = cdm, name = "aspirin")

If we also want the details of these concept IDs we can get these like so.

getDrugIngredientCodes(cdm = cdm, name = "aspirin", withConceptDetails = TRUE)

And if we want codelists for all drug ingredients we can simply omit the name argument and all ingredients will be returned.

ing <- getDrugIngredientCodes(cdm = cdm)
ing$aspirin
ing$diclofenac
ing$celecoxib

Systematic search using CodelistGenerator

CodelistGenerator can also support systematic searches of the vocabulary tables to support codelist development. A little like the process for a systematic review, the idea is that for a specified search strategy, CodelistGenerator will identify a set of concepts that may be relevant, with these then being screened to remove any irrelevant codes by clinical experts.

We can do a simple search for asthma

asthma_codes1 <- getCandidateCodes(
  cdm = cdm,
  keywords = "asthma",
  domains = "Condition"
) 
asthma_codes1 %>% 
  glimpse()

But perhaps we want to exclude certain concepts as part of the search strategy, in this case we can add these like so

asthma_codes2 <- getCandidateCodes(
  cdm = cdm,
  keywords = "asthma",
  exclude = "childhood",
  domains = "Condition"
) 
asthma_codes2 %>% 
  glimpse()

We can compare these two code lists like so

compareCodelists(asthma_codes1, asthma_codes2)

We can then also see non-standard codes these are mapped from, for example here we can see the non-standard ICD10 code that maps to a standard snowmed code for gastrointestinal hemorrhage returned by our search

Gastrointestinal_hemorrhage <- getCandidateCodes(
  cdm = cdm,
  keywords = "Gastrointestinal hemorrhage",
  domains = "Condition"
)
Gastrointestinal_hemorrhage %>% 
  glimpse()

Summarising code use

summariseCodeUse(list("asthma" = asthma_codes1$concept_id),  
                 cdm = cdm) %>% 
  glimpse()

DBI::dbDisconnect(db)

oxford-pharmacoepi/CodelistGenerator documentation built on April 12, 2024, 9:30 a.m.